This document proposes a new approach called a temporal snapshot fact table to analyze insurance data daily over many decades in a space-efficient manner. Traditional solutions like transactional, accumulating, and periodic snapshot fact tables are not feasible due to the large volume of data. The new approach models each fact row as a time interval rather than a point in time. This reduces duplication and allows representing the data using temporal logic and operators. Technical challenges around data integration and cube modeling are addressed through refactoring the source data into a common set of time intervals and modeling the intervals relationally in the cube.